vision-mcp

Convert desktop GUI interaction paths into reusable vision-mcp.yaml maps.

6|2|Updated May 27, 2026
One-click install
npx skills add https://github.com/Haruhiyuki/vision-mcp --skill vision-mcp
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: vision-mcp
Source: https://github.com/Haruhiyuki/vision-mcp/tree/main/skills/vision-mcp
Command: npx skills add https://github.com/Haruhiyuki/vision-mcp --skill vision-mcp

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Vision-MCP reduces the heavy, repetitive cost of desktop GUI tasks by turning each interaction path (screenshots, coordinate estimation, AX/OCR, and click sequences) into a reusable vision-mcp.yaml map that can be replayed with run_workflow. It enables agents to skip exhaustive visual exploration on subsequent runs, cutting initial setup time from minutes to seconds. The approach assumes the agent can perform visual operations and uses a modular map as an amortized automation layer rather than a replacement for existing computer-use skills.

Core Features & Use Cases

  • Path-to-map conversion: every visual operation path is captured and distilled into a reusable map.
  • Workflow reuse: after the first run, subsequent tasks hit run_workflow with near-zero visual cost.
  • Cross-platform support: macOS and Windows compatibility with platform-specific handling and safety checks.
  • Patchable maps: supports patches and repair to keep maps reliable as UI changes.
  • Guided exploration vs task-driven operation: supports both exploration-driven map-building and task-driven execution.

Quick Start

Run vision-mcp workflow with an existing map to execute a desktop task end-to-end.

Frequently Asked Questions about vision-mcp

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate repetitive desktop GUI tasks without recording clicks every time?

Desktop GUI automation can be sped up by converting interaction paths into a reusable YAML map. The map captures screenshots and click sequences, allowing subsequent workflow runs to skip visual exploration and execute instantly with near-zero setup time.

Can I use cross-platform desktop automation maps on both macOS and Windows?

Cross-platform desktop automation maps support both macOS and Windows. The system includes platform-specific handling and safety checks to ensure that reusable interaction paths and click sequences operate reliably across different operating systems.

What is the best way to handle UI changes in automated desktop workflows?

Handling UI changes in automated desktop workflows is done through patchable maps. When an interface updates, you can apply patches and repairs to the existing YAML map to keep the automation reliable without rebuilding the entire interaction path from scratch.

How do I build a desktop automation map from scratch for a new application?

Building a desktop automation map uses guided exploration to capture visual operations like coordinate estimation and OCR. This task-driven process distills each interaction path into a structured YAML schema defining actions and state transitions for later reuse.

Does desktop automation with reusable maps require visual operations on every run?

Desktop automation with reusable maps does not require visual operations on every run. After the first run captures the interaction path, subsequent tasks hit the workflow replay function, cutting the visual cost from minutes to seconds by using the amortized map.

Why does my desktop GUI automation slow down on long-running cross-app workflows?

Desktop GUI automation slows down on long workflows due to exhaustive visual exploration. Converting each interaction path into a reusable YAML map acts as an amortized automation layer, eliminating the repeated visual cost and speeding up cross-app execution.